More than 3 million patient study on type 2 diabetes risk shows potential for more advanced approaches for early detection
new orleans, June 5, 2026 /PRNewswire/ — A new predictive model based on electronic medical records has successfully identified patients who are most at risk of developing type 2 diabetes up to 10 years later. The researchers presented their findings in a public poster session and an electronic poster theater session at the American Diabetes Association’s 2026 Scientific Sessions.® (ADA) in New Orleans.
More than 60% of adults in the United States have risk factors for type 2 diabetes, far more than current diabetes prevention programs can realistically accommodate. Because diabetes often develops gradually over many years without clear warning signs, it can be difficult for health systems and professionals to identify those most at risk who would benefit from prevention programs and early treatment.
This retrospective cohort study included 3,365,464 adults ages 18 to 70 who were treated at Kaiser Permanente in Northern California from 2012 to 2024. Median patient age was 39 years, and 55% of patients were female. The study used a hazard-based superlearning approach that combines multiple survival analysis models to estimate each patient’s risk of developing type 2 diabetes over the next 1, 3, and 10 years. The model used publicly available data such as healthy eating and access to walkable areas, as well as clinical and demographic information routinely collected at clinic visits, such as age, weight, blood sugar (blood sugar) levels, medical history, and medications.
The study found that the incidence of type 2 diabetes was 10.7/1,000 person-years during a median follow-up of 5.4 years. This training model effectively identified adults at high risk for type 2 diabetes with an area under the curve of 0.886 (95% Cl: 0.883 to 0.888). The score of the validation model was 0.883 (95% Cl: 0.88 to 0.886). A 1-year follow-up yielded nearly ideal calibration (average predicted risk 1.03% vs. observed risk 1.01%). At thresholds defining high risk (risk > 1.2%), the model had a sensitivity of 74% and a specificity of 82% over up to 10 years of follow-up.
“These findings demonstrate the potential to advance existing approaches to identifying individuals at risk of developing type 2 diabetes by allowing for earlier and more accurate detection and supporting more targeted and proactive prevention approaches,” said Luis A. Rodriguez, Ph.D., MPH, RD, lead author of the study. “Our model has the potential to create opportunities for clinicians and healthcare organizations to focus prevention efforts on high-risk populations who would benefit most from prevention and treatment and are often missed by traditional screening.”
The authors plan to test this model in a clinical setting to see if it can help increase participation in type 2 diabetes prevention programs and reduce the incidence of diabetes.
Research presentation details
2321-P – Machine learning modeling for T2DM prediction in over 3 million adults [Board No. 2321]
- Luis A. Rodriguez, PhD, MPH, RD
- General poster session
- Saturday, June 6th, 12:30pm – 1:30pm Central Time
- Ernest N. Morial Convention Center, Poster Hall (Hall DE)
2321-P – Machine learning modeling for T2DM prediction in over 3 million adults [Board No. 2321]
- Luis A. Rodriguez, PhD, MPH, RD
- e-Poster Theater – Diabetes Risk and Prediction: One size does not fit all
- Sunday, June 7th, 12:30pm – 1:30pm Central Time
- Ernest N. Morial Convention Center, ePoster Theater B (Hall B1-C)
About the American Diabetes Association’s 2026 Academic Session
The ADA’s 2026 Scientific Sessions, the world’s largest scientific conference focused on diabetes research, prevention, and care, will be held June 5-8 in New Orleans, Louisiana. Thousands of leading physicians, scientists, and medical professionals from around the world are expected to gather in person and virtually to present cutting-edge research, treatment recommendations, and advances toward a cure for diabetes. Attendees will have exclusive access to thousands of unique research presentations and participate in provocative and engaging interactions with leading diabetes experts. Join the Science Sessions conversation on social media using #ADASciSessions.
About the American Diabetes Association
The American Diabetes Association (ADA) is the nation’s leading independent health organization fighting to end diabetes and help people thrive. This year, as ADA celebrates 85 years of advancing discovery and research to prevent, manage, treat, and ultimately cure diabetes, we are not stopping. More than 155 million Americans live with diabetes or prediabetes. We fight for them all through advocacy, program development, and education. To learn more or to get involved, visit diabetes.org or call 1-800-DIABETES (800-342-2383). Join us in the fight on Facebook (American Diabetes Association), Spanish Facebook (Asociación Americana de la Diabetes), LinkedIn (American Diabetes Association), and Instagram (@AmDiabetesAssn). To learn more about how we help everyone living with diabetes, visit X (@AmDiabetesAssn).
Media contact: [email protected]
Source American Diabetes Association
